Publications — Samuel Vaiter

2026

J28 CHANI: Correlation-based Hawkes Aggregation of Neurons with bio-Inspiration.
S. Jaffard, S. Vaiter, and P. Reynaud-Bouret.

Journal of Machine Learning Research. 2026.

arXiv
J27 Bilevel gradient methods and Morse parametric qualification.
J. Bolte, T. Lê, E. Pauwels, and S. Vaiter.

Mathematics of Operations Research. 2026.

DOI arXiv
P5 Fairness-informed Pareto Optimization: An Efficient Bilevel Framework.
S. Tanji, S. Vaiter, and Y. Laguel.

Preprint. 2026.

arXiv
P4 Towards Understanding Steering Strength.
M. Taimeskhanov, S. Vaiter, and D. Garreau.

Preprint. 2026.

arXiv
P3 ZOBA: An Efficient Single-loop Zeroth-order Bilevel Optimization Algorithm.
M. Rando, and S. Vaiter.

Preprint. 2026.

arXiv
P2 Characterizations of inexact proximal operators.
G. Lauga, and S. Vaiter.

Preprint. 2026.

arXiv

2025

J26 Risk Estimate under a Nonstationary Autoregressive Model for Data-Driven Reproduction Number Estimation.
B. Pascal, and S. Vaiter.

Signal Processing. 2025.

DOI arXiv
J25 Deep Equilibrium models for Poisson Imaging Inverse problems via Mirror Descent.
C. Daniele, S. Villa, S. Vaiter, and L. Calatroni.

(accepted to) SIAM Journal on Imaging Sciences. 2025.

arXiv
J24 Geometric and computational hardness of bilevel programming.
J. Bolte, T. Lê, E. Pauwels, and S. Vaiter.

Mathematical Programming. 2025.

DOI arXiv
C21 From Shortcut to Induction Head: How Data Diversity Shapes Algorithm Selection in Transformers.
R. Kawata, Y. Song, A. Bietti, N. Nishikawa, T. Suzuki, S. Vaiter, and D. Wu.

Advances in Neural Information Processing Systems (NeurIPS). 2025.

arXiv
C20 Learning Theory for Kernel Bilevel Optimization.
F. E. Khoury, E. Pauwels, S. Vaiter, and M. Arbel.

Advances in Neural Information Processing Systems (NeurIPS). 2025.

arXiv
C19 Differentiable Generalized Sliced Wasserstein Plans.
L. Chapel, R. Tavenard, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2025.

arXiv
P1 Faster Computation of Entropic Optimal Transport via Stable Low Frequency Modes.
R. Chhaibi, S. Gratton, and S. Vaiter.

Preprint. 2025.

arXiv

2024

J23 A theory of optimal convex regularization for low-dimensional recovery.
Y. Traonmilin, R. Gribonval, and S. Vaiter.

Information and Inference: A Journal of the IMA. 2024.

arXiv
J22 Model identification and local linear convergence of coordinate descent.
Q. Klopfenstein, Q. Bertrand, A. Gramfort, J. Salmon, and S. Vaiter.

Optimization Letters. 2024.

arXiv
J21 Gradient Scarcity with Bilevel Optimization for Graph Learning.
H. Ghanem, S. Vaiter, and N. Keriven.

Transactions on Machine Learning Research. 2024. Featured Certification

arXiv
J20 Convergence of Message Passing Graph Neural Networks with Generic Aggregation On Large Random Graphs.
M. Cordonnier, N. Keriven, N. Tremblay, and S. Vaiter.

Journal of Machine Learning Research. 2024.

arXiv
C18 Provable local learning rule by expert aggregation for a Hawkes network.
S. Jaffard, S. Vaiter, A. Muzy, and P. Reynaud-Bouret.

International Conference on Artificial Intelligence and Statistics (AISTATS). 2024.

PDF arXiv
C17 Derivatives of Stochastic Gradient Descent in parametric optimization.
F. Iutzeler, E. Pauwels, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2024.

arXiv
C16 A Near-Optimal Algorithm for Bilevel Empirical Risk Minimization.
M. Dagréou, T. Moreau, S. Vaiter, and P. Ablin.

International Conference on Artificial Intelligence and Statistics (AISTATS). 2024.

PDF arXiv
W8 How to compute Hessian-vector products?.
M. Dagréou, P. Ablin, S. Vaiter, and T. Moreau.

International Conference on Learning Representations (ICLR) Blogposts Track. 2024.

2023

J19 The Derivatives of Sinkhorn-Knopp Converge.
E. Pauwels, and S. Vaiter.

SIAM Journal on Optimization. 2023.

DOI PDF arXiv
J18 Supervised learning of analysis-sparsity priors with automatic differentiation.
H. Ghanem, J. Salmon, N. Keriven, and S. Vaiter.

IEEE Signal Processing Letters. 2023.

DOI arXiv
J17 The Geometry of Sparse Analysis Regularization.
X. Dupuis, and S. Vaiter.

SIAM Journal on Optimization. 2023.

DOI PDF arXiv HAL
C15 What functions can Graph Neural Networks compute on random graphs? The role of Positional Encoding.
N. Keriven, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2023.

PDF arXiv
C14 On the Robustness of Text Vectorizers.
R. Catellier, S. Vaiter, and D. Garreau.

International Conference on Machine Learning (ICML). 2023.

arXiv
C13 One-step differentiation of iterative algorithms.
J. Bolte, E. Pauwels, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2023. (Spotlight paper)

PDF arXiv
CF5 Borne inférieure de compléxité et algorithme quasi-optimal pour la minimisation de risque empirique bi-niveaux.
M. Dagréou, T. Moreau, S. Vaiter, and P. Ablin.

Colloque Francophone de Traitement du Signal et des Images (GRETSI). 2023.
CF4 Convergence of Message Passing Graph Neural Networks with Generic Aggregation on Random Graphs.
M. Cordonnier, N. Keriven, N. Tremblay, and S. Vaiter.

Colloque Francophone de Traitement du Signal et des Images (GRETSI). 2023.
W7 Convergence of Message Passing Graph Neural Networks with Generic Aggregation on Random Graphs.
M. Cordonnier, N. Keriven, N. Tremblay, and S. Vaiter.

Graph Signal Processing workshop. 2023.

2022

J16 Sparse and Smooth: improved guarantees for Spectral Clustering in the Dynamic Stochastic Block Model.
N. Keriven, and S. Vaiter.

Electronic Journal of Statistics. 2022.

DOI PDF arXiv HAL
J15 Implicit differentiation for fast hyperparameter selection in non-smooth convex learning.
Q. Bertrand, Q. Klopfenstein, M. Massias, M. Blondel, S. Vaiter, A. Gramfort, and J. Salmon.

Journal of Machine Learning Research. 2022.

PDF arXiv
C12 Benchopt: Reproducible, efficient and collaborative optimization benchmarks.
T. Moreau, M. Massias, A. Gramfort, P. Ablin, P.-A. Bannier, B. Charlier, M. Dagréou, T. D. l. Tour, G. Durif, C. F. Dantas, Q. Klopfenstein, J. Larsson, E. Lai, T. Lefort, M. Malezieu, B. Moufad, B. T. Nguyen, A. Rakotomamonjy, Z. Ramzie, J. Salmon, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2022.

PDF arXiv
C11 A framework for bilevel optimization that enables stochastic and global variance reduction algorithms.
M. Dagréou, P. Ablin, S. Vaiter, and T. Moreau.

Advances in Neural Information Processing Systems (NeurIPS). 2022. (Oral paper)

arXiv
C10 Automatic differentiation of nonsmooth iterative algorithms.
J. Bolte, E. Pauwels, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2022.

PDF arXiv
CF3 Algorithmes stochastiques et réduction de variance grâce à un nouveau cadre pour l’optimisation bi-niveaux.
M. Dagr{\'e}ou, P. Ablin, S. Vaiter, and T. Moreau.

Colloque Francophone de Traitement du Signal et des Images (GRETSI). 2022.

2021

J14 Automated data-driven selection of the hyperparameters for Total-Variation based texture segmentation.
B. Pascal, S. Vaiter, N. Pustelnik, and P. Abry.

Journal of Mathematical Imaging and Vision. 2021.

PDF arXiv
J13 Linear Support Vector Regression with Linear Constraints.
Q. Klopfenstein, and S. Vaiter.

Machine Learning. 2021.

DOI arXiv HAL
J12 Block based refitting in ℓ₁₂ sparse regularisation.
C.-A. Deledalle, N. Papadakis, J. Salmon, and S. Vaiter.

Journal of Mathematical Imaging and Vision. 2021.

DOI PDF arXiv HAL
C9 On the Universality of Graph Neural Networks on Large Random Graphs.
N. Keriven, A. Bietti, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2021.

arXiv
T2 From optimization to algorithmic differentiation: a graph detour.
S. Vaiter.

HDR Thesis. 2021.

PDF

2020

J11 Dual Extrapolation for Sparse Generalized Linear Models.
M. Massias, S. Vaiter, A. Gramfort, and J. Salmon.

Journal of Machine Learning Research. 2020.

PDF arXiv HAL
C8 Convergence and Stability of Graph Convolutional Networks on Large Random Graphs.
N. Keriven, A. Bietti, and S. Vaiter.

Advances in Neural Information Processing Systems (NeurIPS). 2020. (Spotlight paper)

PDF arXiv
C7 Implicit differentiation of Lasso-type models for hyperparameter optimization.
Q. Bertrand, Q. Klopfenstein, M. Blondel, S. Vaiter, A. Gramfort, and J. Salmon.

International Conference on Machine Learning (ICML). 2020.

PDF arXiv

2019

J10 Maximal Solutions of Sparse Analysis Regularization.
A. Barbara, A. Jourani, and S. Vaiter.

Journal of Optimization Theory and Applications. 2019.

DOI PDF arXiv HAL
C6 Refitting solutions promoted by ℓ₁₂ sparse analysis regularization with block penalties.
C.-A. Deledalle, N. Papadakis, J. Salmon, and S. Vaiter.

International Conference on Scale Space and Variational Methods in Computer Vision (SSVM). 2019.

arXiv HAL
W6 Exploiting regularity in sparse Generalized Linear Models.
M. Massias, S. Vaiter, A. Gramfort, and J. Salmon.

Signal Processing with Adaptive Sparse Structured Representations (SPARS). 2019.

HAL

2018

J9 Model Consistency of Partly Smooth Regularizers.
S. Vaiter, G. Peyré, and J. Fadili.

IEEE Transactions on Information Theory. 2018.

DOI PDF arXiv HAL
C5 Optimality of 1-norm regularization among weighted 1-norms for sparse recovery: a case study on how to find optimal regularizations.
Y. Traonmilin, and S. Vaiter.

New Computational Methods for Inverse Problems (NCMIP). 2018.

PDF arXiv HAL
W5 Is the 1-norm the best convex sparse regularization?.
Y. Traonmilin, S. Vaiter, and R. Gribonval.

iTWIST. 2018.

PDF arXiv HAL

2017

J8 The Degrees of Freedom of Partly Smooth Regularizers.
S. Vaiter, C.-A. Deledalle, G. Peyré, J. Fadili, and C. Dossal.

Annals of the Institute of Statistical Mathematics. 2017.

DOI PDF arXiv HAL
J7 CLEAR: Covariant LEAst-square Re-fitting with applications to image restoration.
C.-A. Deledalle, N. Papadakis, J. Salmon, and S. Vaiter.

SIAM Journal on Imaging Sciences. 2017.

DOI PDF arXiv HAL
J6 Accelerated Alternating Descent Methods for Dykstra-like problems.
A. Chambolle, P. Tan, and S. Vaiter.

Journal of Mathematical Imaging and Vision. 2017.

DOI PDF HAL
J5 A Sharp Oracle Inequality for Graph-Slope.
P. Bellec, J. Salmon, and S. Vaiter.

Electronic Journal of Statistics. 2017.

DOI PDF arXiv HAL
W4 Characterizing the maximum parameter of the total-variation denoising through the pseudo-inverse of the divergence.
C.-A. Deledalle, N. Papadakis, J. Salmon, and S. Vaiter.

Signal Processing with Adaptive Sparse Structured Representations (SPARS). 2017.

PDF arXiv HAL

2015

J4 Model Selection with Low Complexity Priors.
S. Vaiter, M. Golbabaee, J. Fadili, and G. Peyré.

Information and Inference: A Journal of the IMA. 2015.

DOI PDF arXiv HAL
B1 Low Complexity Regularization of Linear Inverse Problems.
S. Vaiter, G. Peyré, and J. Fadili.

Book chapter in Sampling Theory, a Renaissance. 2015.

DOI PDF arXiv HAL

2014

J3 Stein Unbiased GrAdient estimator of the Risk (SUGAR) for multiple parameter selection.
C.-A. Deledalle, S. Vaiter, J. Fadili, and G. Peyré.

SIAM Journal on Imaging Sciences. 2014.

DOI PDF arXiv HAL
T1 Low Complexity Regularizations of Inverse Problems.
S. Vaiter.

PhD Thesis. 2014.

PDF SITE

2013

J2 Robust Sparse Analysis Regularization.
S. Vaiter, G. Peyré, C. Dossal, and J. Fadili.

IEEE Transactions on Information Theory. 2013.

DOI PDF arXiv HAL
J1 Local Behavior of Sparse Analysis Regularization: Applications to Risk Estimation.
S. Vaiter, C.-A. Deledalle, G. Peyré, C. Dossal, and J. Fadili.

Applied and Computational Harmonic Analysis. 2013.

DOI PDF arXiv HAL
C4 Robust Polyhedral Regularization.
S. Vaiter, G. Peyré, and J. Fadili.

SAMPTA. 2013.

PDF arXiv HAL
C3 Stable Recovery with Analysis Decomposable Priors.
J. Fadili, G. Peyré, S. Vaiter, C.-A. Deledalle, and J. Salmon.

SAMPTA. 2013.

PDF arXiv
CF2 Robustesse au bruit des régularisations polyhédrales.
S. Vaiter, G. Peyré, and J. Fadili.

Colloque Francophone de Traitement du Signal et des Images (GRETSI). 2013.
CF1 Reconstruction Stable par Régularisation Décomposable Analyse.
J. Fadili, G. Peyré, S. Vaiter, C.-A. Deledalle, and J. Salmon.

Colloque Francophone de Traitement du Signal et des Images (GRETSI). 2013.
W3 The degrees of freedom of the group Lasso for a general design.
S. Vaiter, G. Peyré, J. Fadili, C.-A. Deledalle, and C. Dossal.

Signal Processing with Adaptive Sparse Structured Representations (SPARS). 2013.

PDF HAL

2012

C2 Unbiased Risk Estimation for Sparse Analysis Regularization.
C.-A. Deledalle, S. Vaiter, G. Peyré, J. Fadili, and C. Dossal.

International Conference on Image Processing (ICIP). 2012.

DOI PDF HAL
C1 Proximal Splitting Derivatives for Risk Estimation.
C.-A. Deledalle, S. Vaiter, G. Peyré, J. Fadili, and C. Dossal.

New Computational Methods for Inverse Problems (NCMIP). 2012.

DOI PDF HAL
W2 The Degrees of Freedom of the Group Lasso.
S. Vaiter, C.-A. Deledalle, G. Peyré, J. Fadili, and C. Dossal.

International Conference on Machine Learning (ICML) Workshop Sparsity, Dictionaries and Projections in Machine Learning and Signal Processing. 2012.

PDF arXiv HAL
W1 Risk estimation for matrix recovery with spectral regularization.
C.-A. Deledalle, S. Vaiter, G. Peyré, J. Fadili, and C. Dossal.

International Conference on Machine Learning (ICML) Workshop Sparsity, Dictionaries and Projections in Machine Learning and Signal Processing. 2012.

PDF arXiv HAL